Re: Expert systems in forth

On Sun, 5 Jan 2025 17:31:10 +0000, Anton Ertl wrote:

Or, more generally, negative rules.  Then there would be:
>
  bird :- feathers .;
  not bird :- not feathers .;

I'll see how  to do that if possible (I mean 'not' in the goal fact).

I am referring to your rule
>
  bird        :- wings , lay-eggs .;
>
So if you have established that the animal has wings AND lays eggs
(and is not extinct), it's a bird.  With the negative rules one could
also specify
>
  not bird :- not wings .;
  not bird :- not lay-eggs .;
>

Your are right (Logically true), but for the moment I haven't 'not' in
the goal facts.

Until now, I assume: no equivalent to unknown.
Three level logic: yes/no/unknown (true/false/unknown)can be
implemented.
Perhaps, Carnaugh tables can be helpful.

>
Strangely, even though there are a lot of people working on logic in
my school, I have never heard of any work in that direction.  But I
would be very surprised if that was uncharted land.
>

I meant multi-valued logic (three-valued logic).
and when I said Carnaugh tables, I was referring to the possibility to
consider unknown as yes or no given the situation (in the inference
process) but this is a two-valued logic. Perhaps, I was just confused.

- anton

Ahmed
--